General Robotics says its AI software has cut the time needed to set up a new robot from about a month to as little as two hours, taking on engineering work that previously required teams of specialists. 

The Redmond, Washington-based startup, named the approach Auto Engineering. Built into its GRID platform, the system is designed to help engineers connect to robot hardware, develop skills, and fix problems that emerge during testing. 

The two-hour setup claim was reported by GeekWire on Sept. 9. In its technical account, General Robotics describes a broader process taking about four hours to bring a Flexiv robot to its first working skill. Neither figure establishes how quickly a complete factory application could enter regular operation. 

For its laboratory tests, the company used robot arms to handle test tubes and pour liquid into beakers. It says GRID transferred a pouring skill between Flexiv and UR5e arms, reusing work across different hardware. 

Once the initial Flexiv setup was complete, later skills took as little as 10 to 15 minutes. The software carried forward the integration work and fixes from earlier tests, reducing the work needed for the next task. 

People still had a role. A more difficult task, picking up a stirrer and stirring inside a beaker, required human demonstrations to train the system. 

CEO and co-founder Ashish Kapoor stated in an interview with GeekWire that customer projects previously requiring teams of specialists. He described the new approach to help each engineer handle more work. 

General Robotics markets GRID for manufacturing, logistics, energy and government applications, including inspection, goods transport and picking up and placing objects. Its published laboratory results do not show how reliably the software handles those jobs at customer sites. 

Founded in 2023 by former Microsoft researchers, the company has about 50 employees, GeekWire reported. Kapoor said it has roughly a dozen enterprise customers and revenue in the millions of dollars.